i am trying to deal with road detection this last month. So i am working with particle filtering that produces particles, that means possible targets(to my case cars)/points, with hidden information of speed). Further on, i implement spatial+statistical analysis through ArcGIS, using the density of these particles/points as an indicator of where could a road possibly be (through a methodology i follow, is possible to extract a road using the centroids od the polygons - small areas of bigger density of points. I have the slope map and the curvature map of the area. I have also converted everything to rasters and what i would like to do is a percentage probability map that will inform the user that "in this area is impossible what you see to be a target - so it's a false alarm". This will happen to my case in an area that the slope is smaller than 30 degrees of ground declination. Are there any ideas of how could i RELATE SLOPE AND CURVATURE and later to proceed to PERCENTAGE PROBABILITY MAP? I am in a good path to achieve it , but i would appreciate any ideas!Thanks a lot!